Table 2. Best neural network model features for predicting the different outcomes, determined by testing different combinations of activation functions, number of layers, and number of neurons per layer for each data set.
Outcome | First Hidden Layer | Second Hidden Layer | ||
---|---|---|---|---|
Activation Function | Neurons | Activation Function | Neurons | |
FEV1 | Sigmoid | 64 | Sigmoid | 16 |
Smoking Cessation | Hard Sigmoid | 64 | RELU | 32 |
Log Smoking Cessation | Sigmoid | 64 | Sigmoid | 32 |